bar, the distance between them being equal to the original length. The readings are made focusing upon fine platinum wires hung from each end of the specimen and extending downward through and below the furnace.
A simple apparatus for determining thermal expansion is illustrated in Fig. 6, consisting essentially of a container provided with a heating cooling and stirrer. The specimen is placed upon quartz glass supports and a quartz rod is placed upon the upper end. A lever, with a mirror attached to its end, is brought in contact with the quartz rod and the initial and final readings made by means of the telescope and scale. For higher temperatures up to about 300°C. "Crisco" may be used in the bath.
Resistance to Temperature Changes.—Although in general it may be said that the ability of refractories to resist sudden temperature changes is a function of the coefficient of thermal expansion, increasing with the decrease of this con-
FIG. 6.—Testing thermal expansion.
stant, there are many factors affecting this property. Furthermore, the physical changes brought about by the continued heat effect in use may change the character of the refractory very decidedly with time. Thus, a firebrick may resist sudden temperature changes very well in the condition it comes from the kiln but after having been in place in the furnace for some time it may possibly become denser at the surface, through vitrification or the absorption of slag or glass and thus show a decided tendency to spall. The specimens therefore should be fired to a temperature at or above that of their use before being tested. With some kinds of refractories their tendency to spall when heated or cooled quickly constitutes a serious draw-back to their use.
No standardized tests for the estimation of this quality are available at the present time and the practices vary widely. The specimens are heated to temperatures varying from 800 to 1,000°C. and are subjected to treatments consisting in quenching the test pieces in cold water, either by complete or partial immersion or in subjecting them to cooling in an air blast. After such treatment, varying from 1 to 5 min., the specimens are replaced in the heating furnace and the operation repeated a specified number of times or until failure occurs. The criteria for comparison may be the number of treatments withstood, the weight of material broken off or the ratio between the original and final strength, usually the modulus of rupture. The last procedure would seem to be the most rational.
Resistance to Slagging Action.—According to the porosity of the refractories, the temperature, viscosity and the chemical nature of the slag or glass the absorp-